On the Characterization of a Class of Fisher-Consistent Loss Functions and its Application to Boosting.
Matey NeykovJun S. LiuTianxi CaiPublished in: J. Mach. Learn. Res. (2016)
Keyphrases
- loss function
- boosting algorithms
- boosting framework
- gradient boosting
- reproducing kernel hilbert space
- partially labeled data
- pairwise
- bregman divergences
- loss minimization
- logistic regression
- squared error
- learning to rank
- support vector
- update rules
- risk minimization
- hinge loss
- learning theory
- cost sensitive
- square loss
- convex loss functions